RESEARCH ARTICLE

Constitutive activation of kappa receptors at ventral tegmental area inhibitory synapses following acute stress Abigail M Polter1†, Kelsey Barcomb1, Rudy W Chen1, Paige M Dingess2,3, Nicholas M Graziane1‡, Travis E Brown2,3, Julie A Kauer1*

1Department of Molecular Pharmacology, Physiology and Biotechnology, Brown University, Providence, United States; 2Neuroscience Program, University of Wyoming, Laramie, United States; 3University of Wyoming, School of Pharmacy, Laramie, United States

Abstract Stressful experiences potently activate kappa opioid receptors (kORs). kORs in the ventral tegmental area regulate multiple aspects of dopaminergic and non-dopaminergic cell function. Here we show that at GABAergic synapses on rat VTA dopamine neurons, a single exposure to a brief cold-water swim stress induces prolonged activation of kORs. This is mediated *For correspondence: by activation of the receptor during the stressor followed by a persistent, ligand-independent [email protected] constitutive activation of the kOR itself. This lasting change in function is not seen at kORs at neighboring excitatory synapses, suggesting distinct time courses and mechanisms of regulation of Present address: †Department different subsets of kORs. We also provide evidence that constitutive activity of kORs governs the of Pharmacology and Physiology, The George Washington prolonged reinstatement to cocaine-seeking observed after cold water swim stress. Together, our University School of Medicine studies indicate that stress-induced constitutive activation is a novel mechanism of kOR regulation and Health Science, Washington, that plays a critical role in reinstatement of drug seeking. United States; ‡Department of DOI: 10.7554/eLife.23785.001 Neuroscience, University of Pittsburgh, Pittsburgh, United States

Competing interests: The Introduction authors declare that no Stress has long been known to be a precipitating factor for the abuse of addictive drugs. Animal competing interests exist. models have shown that acute and repeated stressors can escalate intake of addictive substances (Piazza et al., 1990; Ramsey and Van Ree, 1993; Goeders and Guerin, 1994; Shaham and Stew- Funding: See page 17 art, 1994; Haney et al., 1995), and that acute stress can reinstate drug seeking in animals that have Received: 30 November 2016 undergone extinction training (Shaham et al., 1994, 1995; Conrad et al., 2010; Mantsch et al., Accepted: 13 March 2017 2016). In recent years, dopaminergic neurons of the VTA have emerged as a significant locus for the Published: 12 April 2017 overlapping effects of drugs of abuse and stress (Polter and Kauer, 2014). Synaptic inputs, by shap- Reviewing editor: Lisa M ing the activity of these neurons, are poised to play an important role in drug seeking. Both acute Monteggia, University of Texas stress and exposure to drugs of abuse induce a concomitant potentiation of excitatory synapses and Southwestern Medical Center, loss of long term potentiation of inhibitory synapses (Ungless et al., 2001; Saal et al., 2003; United States Kauer and Malenka, 2007; Nugent et al., 2007; Chen et al., 2008; Niehaus et al., 2010; Polter and Kauer, 2014). Understanding how these synapses are altered by stress will provide key Copyright Polter et al. This insights into stress-induced drug seeking and provide targets for treating substance use disorders. article is distributed under the A major mediator of stress-induced changes in inhibitory VTA synapses is the /kappa terms of the Creative Commons Attribution License, which (kOR) system. kORs, and their endogenous ligand, dynorphin, are found throughout permits unrestricted use and the brain and have been highly associated with stressful, aversive, and dysphoric experiences redistribution provided that the (Bruchas et al., 2010; Wee and Koob, 2010; Van’t Veer and Carlezon, 2013; Crowley and Kash, original author and source are 2015). Within the VTA, kORs have a range of physiological effects. kORs decrease the firing rate of credited. dopamine neurons through activation of GIRK channels (Margolis et al., 2003, 2006), inhibit

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 1 of 21 Research article Neuroscience

eLife digest People who are recovering from drug addiction are more vulnerable to cravings and relapse when under stress. This ability of stress to boost drug relapse can also be shown in animals previously exposed to addictive drugs. Rats can learn to press a lever to administer themselves a dose of cocaine and, during withdrawal, rats previously exposed to the drug will press the lever more often if they are stressed. Indeed, just a few minutes of stress is enough to increase lever pressing for several days. Stress and addictive drugs both act on a region of the brain called the ventral tegmental area, or VTA, which is part of the brain’s reward system. Stress indirectly increases the activity of the VTA. It does so by activating a protein on the surface of VTA neurons called the kappa opioid receptor (kOR for short). Previous studies revealed that five minutes of stress increases the activity of kORs in the VTA of rats for five days. Conversely, blocking kORs stopped stressed rats from pressing the lever more often for cocaine. Together, these findings suggested that activating kORs in the VTA contributes to stress-induced drug relapse. Polter et al. have now discovered how stress activates kORs. It turns out that stressful or unpleasant experiences cause the brain to produce a protein called dynorphin, which binds to and activates the kORs. After a stressful event, the receptors are said to have become constitutively active, and blocking this constitutive activity prevents stress from inducing drug-seeking behavior. Polter et al. show that binding of dynorphin is needed to change the shape of the receptors so that they remain active even after dynorphin has detached, but it is likely that other molecules are also involved. This is the first study to show a link between stress, constitutive activation of kORs, and drug relapse. The next step is to work out why this process occurs on only some and not all occasions when the brain releases dynorphin, and why only certain kORs in the VTA respond in this way. Whether constitutive kOR activity drives stress-related craving in people with a history of drug abuse and how to halt these cravings also remain to be determined. DOI: 10.7554/eLife.23785.002

excitatory synaptic transmission onto both dopaminergic and non-dopaminergic VTA neurons (Margolis et al., 2005), reduce inhibitory synaptic transmission in a subset of dopamine neurons (Ford et al., 2006) and inhibit somatodendritic dopaminergic IPSCs (Ford et al., 2007). VTA kORs also can control the interactions between stress and reward. Our previous work identified a form of

stress-sensitive synaptic plasticity at inhibitory synapses on VTA dopamine neurons (LTPGABA;

Nugent et al., 2007, 2009; Niehaus et al., 2010). LTPGABA is induced via activation of nitric oxide synthase in the dopamine neuron, leading to nitric oxide (NO) release, and enhancement of GABA release through cGMP signaling (Nugent et al., 2007, 2009).

More recently, we showed that acute stress blocks LTPGABA through activation of kORs, and that preventing this activation via intra-VTA administration of the kOR antagonist, nor- (norBNI), prevents stress-induced reinstatement of cocaine-seeking (Graziane et al., 2013). Remark-

ably, a single exposure to stress leads to a loss of LTPGABA that lasts for at least five days and is mediated by persistent activation of VTA kORs (Polter et al., 2014). We have also shown that treat- ment with the kOR antagonist after stress can rescue stress-induced reinstatement. These studies highlight the importance of kOR-mediated regulation of LTP at GABAergic synapses in stress- induced drug seeking and underscore the need to better understand the mechanism of this unique and persistent regulation. In the present study, we have now identified the mechanism by which activation of kORs and sup-

pression of LTPGABA in the VTA is maintained for multiple days after an acute, severe stressor. We

present evidence that stress blocks LTPGABA by inducing constitutive activation of kORs at VTA inhibitory synapses rather than through persistent increases in dynorphin release. This constitutive activity is likely to be triggered initially by signaling through the endogenous ligand dynorphin, but then is persistently maintained independently of dynorphin release. In parallel, we find that the per- sistent drug-seeking induced by a single exposure to acute stress is also dependent on constitutive activity of kORs. Our results reveal a novel mechanism of experience-dependent regulation of kOR

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 2 of 21 Research article Neuroscience function, and emphasize the essential role of kORs in mediating stress-induced changes in synaptic plasticity and drug-seeking behavior.

Results

JNK-dependent rescue of LTPGABA by acute norBNI As previously shown, bath application of the nitric oxide donor SNAP potentiates GABAergic synap- ses on dopamine neurons in the VTA, similarly to high-frequency stimulation of VTA afferents; this potentiation is blocked by single exposure to multiple drugs of abuse or acute cold-water swim

stress (LTPGABA; Nugent et al., 2007; Niehaus et al., 2010; Graziane et al., 2013; Polter et al., 2014; Figure 1A–B). Our recent studies indicate that blocking kORs with norNBI prevents and

reverses the effects of acute stress on LTPGABA, even when administered several days after stress (Graziane et al., 2013; Polter et al., 2014). We therefore investigated whether stress-induced, per- sistent activation of kORs could be detected ex vivo in the midbrain slice. We subjected rats to acute cold water forced swim stress and prepared midbrain slices 24 hr later. If after stress, kORs in the VTA are persistently signaling in vitro, we reasoned that bath-applied norBNI could be used to

rescue SNAP-induced LTPGABA. Bath application of norBNI (100 nM) indeed allowed us to elicit NO-

dependent LTPGABA in slices from stressed animals (Figure 1E), indicating that stress-induced activ- ity of kORs in the VTA persists through brain slice preparation and recovery. It seemed unlikely that sufficient endogenous dynorphin could be released tonically from the denervated brain slices to

maintain a block of LTPGABA in vitro. We therefore next sought to establish the mechanism by which norBNI rescued this plasticity. In addition to competing with agonists at the kOR agonist binding site, norBNI acts as an inverse or collateral agonist, and its interactions with the kOR can non-com- petitively inhibit further activity of kORs via activation of the JNK signaling cascade (Bruchas et al.,

2007; Melief et al., 2010, 2011). We hypothesized that the rescue of LTPGABA by norBNI might also occur non-competitively via JNK signaling (Figure 1C). Slices were treated with the JNK inhibitor SP600125 (20 mM) for 10 min prior to bath application of norBNI (Figure 1D). In contrast to the robust SNAP-induced potentiation observed in slices treated with norBNI alone, we found that

LTPGABA remained blocked in slices pretreated with SP600125 (Figure 1F–H). Importantly, bath

application of SP600125 did not interfere with expression of LTPGABA in slices from naı¨ve animals or

the loss of LTPGABA in slices from stressed animals (Figure 1—figure supplement 1A–B). Therefore,

JNK activity has no role in LTPGABA induction or in the block of this plasticity by kORs, but is

required for norBNI to rescue LTPGABA following stress.

LTPGABA is not rescued by a neutral antagonist Our data indicate that following stress, kOR activation persists even in the brain slice, and is rescued in a JNK-dependent manner. This suggests that non-competitive actions of norBNI, rather than its

block of dynorphin binding, are relevant to the loss of LTPGABA. To test this hypothesis further, we again utilized pharmacological tools in slices from stressed animals. We treated such slices with either norBNI or 6b-naltrexol, a neutral antagonist that only inhibits agonist-stimulated kOR activity

(Figure 2A–B; Wang et al., 2007). If norBNI rescues LTPGABA only because it can activate JNK sig- naling, we would predict that a neutral antagonist that only inhibits kOR agonist binding would be

ineffective (Figure 2A, Wang et al., 2007). While norBNI treatment rescued LTPGABA, bath applica-

tion of the neutral antagonist did not reverse the stress-induced block of LTPGABA (Figure 2C–F). Bath application of 6b-naltrexol was sufficient to prevent depression of EPSCs onto VTA dopamine neurons induced by the kOR agonist U50488 (Figure 2—figure supplement 1B, Margolis et al., 2005), indicating that this concentration of the drug is sufficient to block kORs in the VTA slice. 6b- naltrexol did not have any effects on basal inhibitory transmission in slices from stressed or naı¨ve rats (Figure 2—figure supplement 1A). These results show that a kOR competitive antagonist cannot

effectively rescue LTPGABA following stress, and suggest that the persistent block of LTPGABA is main- tained by constitutive activation of kORs in the VTA rather than a prolonged increase in dynorphin release.

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 3 of 21 Research article Neuroscience

a b 2.0 2.5 SNAP 2.0 1.5

1.5 1.0 * 1.0

TP Magnitude TP 0.5 0.5 Control L Stress 0.0 0.0 Norm. IPSC Amplitude IPSC Norm. 0 10 20 30 40 control stress Time ( min) c nor d BNI Prepare norBNI Stress Slices SP600125 record

κOR JNK SP600125 24 h 1 h 10m

e norBNI f SP600125 +norBNI

800 norBNI 800 SP 600125 + norBNI SNAP SNAP 600 600

400 400

200 200 IPSC Amplitude(pA) IPSC IPSC Amplitude(pA) IPSC 0 0 0 10 20 30 40 0 10 20 30 40 Time (min) Time (min) g h 2.0 2.5 SNAP 2.0 1.5 * 1.5 1.0

1.0 Magnitude

norBNI TP 0.5 0.5 L SP600125+ norBNI 0.0 0.0 Norm. IPSC Amplitude IPSC Norm. 0 10 20 30 40 norBNI SP600125 Time (min) + norBNI

Figure 1. norBNI rescues LTPGABA through activation of JNK. (A) Summary data showing the blockade of LTPGABA after stress. (B) Comparison of the magnitude of LTPGABA10–15 min after SNAP application. (IPSC amplitudes, control: 140 ± 5% of baseline values, n = 13; stress: 94 ± 11% of baseline values, n = 6; unpaired t-test, *p=0.0005. (C) Schematic of norBNI’s competitive and non-competitive inhibition of kOR signaling. (D) Experimental design. (E) Representative single experiment showing that bath application of norBNI (100 nM) rescues LTPGABA in a slice prepared 24 hr after stress. (F)

Representative single experiment from a slice prepared 24 hr after stress showing that norBNI does not rescue LTPGABA in the presence of the JNK inhibitor SP600125 (20 mM). (G) Summary data from both groups. (H) Comparison of the magnitude of LTPGABA10–15 min after SNAP application. (IPSC amplitudes, norBNI only: 139 ± 7% of baseline values, n = 6; norBNI+SP600125: 106 ± 9% of baseline values, n = 11; unpaired t-test, *p=0.029.) Insets for this and all figures: IPSCs before (black trace, control) and 15 min after drug application (red trace, SNAP, 400 mM). Scale bars: 20 ms, 100 pA. Insets are averages of 12 IPSCs. DOI: 10.7554/eLife.23785.003 The following figure supplement is available for figure 1:

Figure supplement 1. Inhibition of JNK does not affect LTPGABA or its block by stress in the absence of norBNI. DOI: 10.7554/eLife.23785.004

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 4 of 21 Research article Neuroscience

a b nor norBNI/ BNI 6β-nal Prepare 6β-nal Stress Slices record κ κ 24 h 1 h

c stress + norBNI d stress + 6 β− nal

norBNISNAP (400 uM) 6β− naltrexol 1000 SNAP 1000 SNAP 800 800 600 600 400 400 200 200 IPSC Amplitude(pA) IPSC

IPSC Amplitude(pA) IPSC 0 0 0 10 20 30 40 0 10 20 30 40 Time ( min) Time (min) e f 2.0 2.5 SNAP 2.0 1.5 * Amplitude 1.5 1.0 1.0 0.5 0.5 norBNI Magnitude LTP 0.0 6β-nal 0.0 Norm. IPSC Norm. 0 10 20 30 40 norBNI 6β-nal Time ( min)

Figure 2. The neutral antagonist 6b-naltrexol fails to rescue LTPGABA in slices from stressed animals. (A) Schematic of norBNI and 6b-naltrexol inhibition of kOR signaling. (B) Experimental design. (C) Representative experiment showing that bath application of norBNI (100 nM) rescues LTPGABA in a slice prepared 24 hr after stress. (D) Representative experiment from a cell 24 hr after stress showing that 6b-naltrexol (10 mM) fails to rescue

LTPGABA.(E) Summary data from both groups. (F) Comparison of the magnitude of LTPGABA10–15 min after SNAP application. (IPSC amplitudes, norBNI: 141 ± 20% of baseline values, n = 10; 6b-naltrexol: 100 ± 8% of baseline values, n = 10; unpaired t-test, *p=0.048). DOI: 10.7554/eLife.23785.005 The following figure supplement is available for figure 2: Figure supplement 1. 6b-naltrexol does not affect basal inhibitory synaptic transmission but does block kORs. DOI: 10.7554/eLife.23785.006

Transient kOR activation leads to persistent kOR activity How might acute stress cause constitutive activation of kORs? While the results of our slice experi- ments rule out a requirement for elevated dynorphin in maintaining persistent activity of VTA kORs following stress, dynorphin release during or immediately following stress may be needed to trigger a change in the receptor leading to prolonged constitutive activation. If this model is correct, pre-

venting binding of dynorphin to the kOR during stress would prevent the loss of LTPGABA. However

after stress, when the block of LTPGABA is no longer dynorphin-dependent, preventing dynorphin

binding would not rescue LTPGABA. To test this idea, we treated animals with the competitive antag- onist 6b–naltrexol either 30 min before or one day after FSS (Figure 3A). Consistent with our

hypothesis, cells from animals treated with 6b-naltrexol before stress exhibited LTPGABA, while those

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 5 of 21 Research article Neuroscience

a Prepare b Vehicle + stress 6β-nal stress Slices

6β− nal 1000 SNAP pre-stress: 30 min 24 h 800 600

Prepare 400 stress 6β-nal Slices 200 6β− nal

post-stress: Amplitude(pA) IPSC 0 24 h 30 min 0 10 20 30 40 Time (min)

c 6β− nal pre-stress d 6β− nal post-stress

1000 SNAP 1000 SNAP 800 800 600 600 400 400 200 200 IPSC Amplitude(pA) IPSC 0 Amplitude(pA) IPSC 0 0 10 20 30 40 0 10 20 30 40 Time (min) Time (min)

e f 2.0 2.5 SNAP *

1.5 2.0 ns 1.5 1.0 1.0 6β-nal pre 0.5

0.5

6β-nal post Magnitude LTP Vehicle 0.0 0.0 Norm. IPSC Amplitude IPSC Norm. 0 10 20 30 40 6β-nal 6β-nal Veh Time ( min) pre post

Figure 3. 6b-naltrexol rescues LTPGABA when administered pre-stress, but not post-stress. (A) Experimental design. (B) Representative experiment showing that a cell from a vehicle-treated stressed animal does not exhibit LTPGABA.(C) Representative experiment showing that a cell from an animal treated with 6b-naltrexol (10 mg/kg) 30 min pre-stress exhibits LTPGABA.(D) Representative experiment showing that a cell from an animal treated with

6b-naltrexol 24 hr post-stress does not exhibit LTPGABA.(E) Summary data showing compiled data from all groups. (F) Comparison of the magnitude of

LTPGABA 10–15 min after SNAP application. (1-way ANOVA followed by Dunnett’s multiple comparison test. F2, 30=4.231,p=0.024. IPSC amplitudes, 6b- naltrexol pre-stress: 136 ± 12% of baseline values, n = 12, p<0.05 from vehicle; 6b-naltrexol post-stress: 100 ± 9% of baseline values, n = 11, n.s. from vehicle; vehicle+stress: 102 ± 8% of baseline values, n = 10). DOI: 10.7554/eLife.23785.007

treated one day after stress did not, similarly to the vehicle-treated animals (Figure 3B-F). In con- trast, our previous studies have shown that treating rats with norBNI at the same time point after

stress (one day) rescues LTPGABA (Polter et al., 2014). These data strongly support the idea that the

persistent block of LTPGABA following acute swim stress is mediated by dynorphin-dependent

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 6 of 21 Research article Neuroscience activation of the kOR followed by a transition to dynorphin-independent constitutive activity of the receptor. To investigate whether a brief activation of kORs is sufficient to produce persistently activated

kORs, we treated rats with a single dose of the kOR agonist, U50488, and measured LTPGABA at var- ious time points after injection (Figure 4A). Upon injection, U50488 rapidly enters the CNS and is metabolized and undetectable in the brain by 24 hr after administration (Russell et al., 2014), and

a b Saline

800 U50488/ Prepare SNAPSNAP 400 µM vehicle Slices 600

400 1 day or 5 days 200

IPSC Amplitude(pA) IPSC 0 0 10 20 30 40 Time (min)

c 1 d. post U50488 d 5 d. post U50488

800 SNAP 800 SNAP 600 600

400 400

200 200

IPSC Amplitude(pA) IPSC 0

0 Amplitude(pA) IPSC 0 10 20 30 40 0 10 20 30 40 Time (min) Time (min)

e f 2.0 * 2.5 SNAP 1.5 2.0 1.5 1.0 1.0 Magnitude 0.5 0.5 Saline U50488 5 mg/kg 1 day LTP 0.0 U50488 5 mg/kg 5 day 0.0 Norm. IPSC Amplitude IPSCNorm. 0 10 20 30 40 Sal U50 U50 Time ( min) 1 d. 5 d.

Figure 4. Single treatment with a kOR agonist leads to prolonged blockade of LTPGABA.(A) Experimental design. (B) Representative experiment showing that a cell from a saline-treated animal exhibits LTPGABA.(C) Representative single experiment showing a cell prepared 24 hr after a single treatment with U50488 (5 mg/kg) does not exhibit LTPGABA.(D) Representative experiment showing that a cell prepared five days after a single treatment with U50488 does not exhibit LTPGABA.(E) Summary data from all groups. (F) Comparison of the magnitude of LTPGABA10–15 min after SNAP application. (1-way ANOVA followed by Dunnett’s multiple comparison test. F2, 27=12.21, p=0.0002. IPSC amplitudes, Saline: 137 ± 6% of baseline values, n = 11; U50488 1 day: 108 ± 6% of baseline values, n = 9, p<0.05 vs. saline; U50488 5 days: 100 ± 5% of baseline values, n = 10, p<0.05 vs. saline). DOI: 10.7554/eLife.23785.008

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 7 of 21 Research article Neuroscience we therefore expect that within our experimental time frame, U50488 was no longer occupying the kOR. In neurons from saline-treated animals, bath application of SNAP robustly potentiated IPSCs

(Figure 4B). In contrast, SNAP was unable to elicit LTPGABA in neurons from rats either one or five days after U50488 administration (Figure 4C–F). Notably, this time course closely mirrors that of the

in vivo block of LTPGABA following acute stress (Polter et al., 2014). Specificity to inhibitory VTA synapses We next addressed the question of whether kORs at other brain synapses are also persistently acti- vated after acute stress. Bath application of the kOR agonist U69593 has been reported to depress

the amplitude of glutamatergic EPSCs in both VTA Ih positive (presumptive dopamine neurons) and

Ih negative (presumptive non-dopamine neurons), and norBNI reverses this depression (Margolis et al., 2005). Therefore if kORs at excitatory synapses become constitutively activated after swim stress, reducing their activity with norBNI should be detectable as potentiation of excit- atory VTA synapses. To test this, we prepared VTA slices 24 hr after FSS. We recorded EPSCs from

Ih positive and Ih negative neurons from both stressed and unstressed animals and bath-applied

norBNI. NorBNI had no effect on EPSC amplitude in Ih-positive neurons in slices from either naı¨ve or

stressed animals (Figure 5A–C), and norBNI did not increase EPSC amplitudes in VTA Ih-negative neurons in slices from either naı¨ve or stressed animals (Figure 5D–F). Therefore, the persistent con- stitutive kOR activation we observe at GABAergic synapses after acute stress does not occur at all kORs, even within the VTA.

Constitutive activity of kORs is required for prolonged stress-induced cocaine seeking Numerous studies from our lab and others have shown a close association between kOR activation and stress-induced drug-seeking behavior (McLaughlin et al., 2003; Redila and Chavkin, 2008; Land et al., 2009; Wee and Koob, 2010; Graziane et al., 2013; Zhou et al., 2013; Polter et al., 2014). We recently reported that blocking kORs with norBNI reverses the modest but prolonged reinstatement of cocaine-seeking induced by swim stress (Conrad et al., 2010; Graziane et al., 2013). This rescue is seen even when norBNI is administered two hours after stress (Polter et al., 2014). These findings are consistent with the hypothesis that reinstatement of cocaine-seeking after

swim stress requires activation of VTA kORs and suppression of LTPGABA. Having now shown that

the blockade of LTPGABA by swim stress is dependent on constitutive activity of kORs, we next tested whether reinstatement of cocaine seeking is similarly dependent on constitutively active kORs. Rats were trained to self-administer cocaine for a minimum of 10 days. Animals then underwent extinction training, and after the final extinction session, they were subjected to forced swim stress, and then returned to their home cages. Twenty-four hours after stress, one group of animals was treated with norBNI and a second group with saline. A third group was treated with 6b–naltrexol 2 days after stress and 60 min prior to reinstatement testing (Figure 6A). Due to the differing pharma- cokinetic profiles of norBNI and 6b–naltrexol, time of administration was varied to optimize block of the kOR during the reinstatement test and to ensure that all animals were tested for reinstatement at the same time point (Endoh et al., 1992; Raehal et al., 2005); thus, all animals were tested for reinstatement 48 hr after stress. As previously shown, after acute stress, vehicle-treated animals showed a significant elevation of lever pressing compared to the final extinction session (Figure 6B). Although the reinstatement was modest, this was measured two full days after the stress, demonstrating the prolonged increase in cocaine-seeking (Conrad et al., 2010). In contrast, animals given norBNI 24 hr post-stress did not increase their lever pressing two days after stress (Figure 6B). Moreover, the neutral antagonist 6b- naltrexol did not prevent reinstatement, as 6b–naltrexol treated animals significantly increased lever pressing compared to the final extinction session (Figure 6B). These data suggest that while persis- tent activation of kORs underlies the prolonged reinstatement induced by swim stress, this is medi- ated by constitutively active receptors rather than by long-term increases in the level of the endogenous ligand dynorphin.

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 8 of 21 Research article Neuroscience

a Ih+ neuron b Ih+ neuron c control stress

400 norBNI 400 norBNI 2.0 norBNI norBNI 100nM 300 300 1.5

200 200 1.0 Amplitude(pA) Amplitude(pA) 100 100 0.5 control stress EPSC EPSC 0 0 0.0 0 10 20 30 0 10 20 30 Amplitude EPSCNorm. 0 10 20 30 Time (min) Time (min) Time ( min)

d Ih- neuron e Ih- neuron f control stress

500 norBNInorBNI 500 norBNI 2.0 norBNI 400 400 1.5 300 300 1.0 200 200 Amplitude(pA) 100 100 0.5 control stress EPSC Amplitude(pA) EPSC 0 EPSC 0 0.0 0 10 20 30 0 10 20 30 Amplitude EPSCNorm. 0 10 20 30 Time (min) Time (min) Time (min)

Figure 5. kORs at VTA excitatory synapses are not constitutively activated by stress. (A) Representative experiment showing that norBNI (100 nM) does not potentiate excitatory synapses on Ih+ VTA neurons in a slice prepared from a control animal. (B) Representative experiment showing that norBNI does not potentiate excitatory synapses on Ih+ VTA neurons in a slice prepared from a stressed animal. (C) Summary data from Ih+ neurons. No significant difference in IPSC amplitude 10–15 min after norBNI application (t-test p=0.81 IPSC amplitudes, control: 94 ± 2% of baseline values, n = 5; stressed: 92 ± 6% of baseline values, n = 6). (D) Representative experiment showing that norBNI does not potentiate excitatory synapses on IhÀ VTA neurons in a slice prepared from a control animal. (E) Representative single experiment showing that norBNI does not potentiate excitatory synapses on IhÀ VTA neurons in a slice prepared from a stressed animals. (F) Summary data from IhÀ neurons. No significant difference in IPSC amplitude 10–15 min after norBNI application (t-test p=0.49 IPSC amplitudes, control: 112 ± 2% of baseline values, n = 5; stressed: 110 ± 4% of baseline values, n = 5). DOI: 10.7554/eLife.23785.009

Discussion

Acute stress causes a loss of plasticity at VTA GABAA synapses that lasts for days and is caused by persistent activation of kORs (Graziane et al., 2013; Polter et al., 2014). This activation could be caused either by a prolonged increase in dynorphin or by an increase in constitutive activity of kORs. In this study, our data support the latter mechanism: a single exposure to an acute stressor causes a lasting constitutive activation of VTA kORs that suppresses plasticity at inhibitory synapses corre- lated with stress-induced drug-seeking (Figure 7). While previous studies have demonstrated consti- tutive activity of kORs in cultured cells and in the rat brain (Wang et al., 2007; Sirohi and Walker, 2015), ours is the first demonstration of experience-induced changes in constitutive activity of these receptors. This represents a novel mechanism of regulation by acute stress of the dynorphin-kOR system, and sheds new light on signaling pathways involved in reinstatement of drug seeking.

Constitutive activation of VTA kORs It is widely accepted that GPCRs can adopt agonist-independent conformations that are constitu- tively active (Seifert and Wenzel-Seifert, 2002; Sade´e et al., 2005;Young et al., 2013 Meye et al.,

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 9 of 21 Research article Neuroscience a Saline Train saline Extinction stress RI ~ 10 d ~ 20 d 24 h 24 h test norBNI norBNI Train Extinction stress RI ~ 10 d ~ 20 d 24 h 24 h test 6β− Naltrexol Train Extinction 6βNal RI stress ~ 10 d ~ 20 d 47 h 1 h test

b saline norBNI 6β -naltrexol 20 * 20 20 15 15 15 *

10 10 10

5 5 5 LeverPresses LeverPresses LeverPresses

0 0 0

final extinction session reinstatement

Figure 6. Post-stress rescue of reinstatement by norBNI but not 6b-naltrexol. (A) Experimental design. (B) Lever pressing during the final extinction session (white bar) and reinstatement session (colored bar). Saline (black): last extinction session: 6.4 ± 1.7 lever presses; reinstatement session: 13.8 ± 2.4 lever presses; n = 8, *p=0.011, paired t-test. norBNI (green): last extinction session: 4.8 ± 1.4 lever presses; reinstatement session: 5.2 ± 1.1 lever presses; n = 6, p=0.76, paired t-test. 6b-naltrexol (blue): last extinction session: 5.7 ± 1.8 lever presses; reinstatement session: 11.2 ± 3.4 lever presses; n = 11, *p=0.033, paired t-test. DOI: 10.7554/eLife.23785.010

2014). In addition to kORs, the other members of the opioid receptor subfamily, mOR and dOR, have both been shown to exhibit constitutive activity (Wang et al., 1994, 2004; Chiu et al., 1996; Wang et al., 1999; Liu and Prather, 2001; Wang et al., 2007; Corder et al., 2013). kORs them- selves have been shown to exhibit constitutive activity, both in heterologous expression systems (Becker et al., 1999; Wang et al., 2007) and in the rat PFC (Sirohi and Walker, 2015). A decrease

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 10 of 21 Research article Neuroscience

a during stress

norBNI DYN 6βNal LTP GABA κOR

b after stress

κOR LTP GABA

c after stress+norBNI

norBNI

JNK κOR LTPLTPGABA

Figure 7. Constitutive activation of kORs by stress. (A) During stress, dynorphin binding to the kOR triggers a shift to a constitutively active state. By blocking dynorphin binding, both norBNI and 6b-naltrexol prevent the loss of

LTPGABA during this time. (B) After stress, the block of LTPGABA is maintained by constitutive activity of kORs and is

no longer dependent on dynorphin binding. (C) norBNI reverses the stress-induced block of LTPGABA by activating the JNK signaling pathway which non-competitively reduces kOR activity. DOI: 10.7554/eLife.23785.011

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 11 of 21 Research article Neuroscience in fear and anxiety behaviors in rats after acute footshock has also been reported that is reduced by post-shock norBNI, suggesting the possibility of constitutive kOR activation, although a high dose was required and the effect of norBNI was not compared to a neutral antagonist (Rogala et al., 2012). Very little is known, however, about the processes that regulate transitions between constitu- tively active and inactive states, presumably representing distinct receptor conformations (Seifert and Wenzel-Seifert, 2002; Sade´e et al., 2005). We present two critical pieces of data indicating that stress induces constitutive activity of kORs. First, brief application (~15 min) of a kOR inverse agonist to VTA slices from stressed rats rescues

LTPGABA in a JNK-dependent fashion. Second, the neutral antagonist does not rescue LTPGABA. Sig- naling through the JNK pathway is thought to be responsible for the long-lasting non-competitive inhibition of the kOR (Bruchas et al., 2007; Melief et al., 2010, 2011). In our experiments, the

requirement of JNK for norBNI to rescue LTPGABA is evidence that norBNI acts through non-compet- itive means, and suggests that the persistent activation of the kOR following stress does not require continuous receptor binding by ligand. Importantly, inhibition of JNK signaling alone did not prevent

LTPGABA induction in slices from naı¨ve animals, nor did it restore LTPGABA in slices from stressed ani- mals, indicating that the role of JNK is limited to inhibition of the receptor by norBNI. The failure of

the JNK inhibitor to rescue LTPGABA indicates that inhibition of LTPGABA by kORs is not mediated by the JNK pathway, but instead most likely through one of the other pathways downstream of kORs,

such as the p38 or ERK MAPK pathways, or through activation of Gai (Bruchas and Chavkin, 2010; In˜iguez et al., 2010; Ehrich et al., 2015).

The inability of the neutral antagonist, 6b–naltrexol, to rescue LTPGABA is consistent with stress promoting constitutive kOR activity at inhibitory VTA synapses. NorBNI, through activation of JNK, reduces the signaling capacity of the kOR regardless of whether this occurs through constitutive activity or increased dynorphin binding. In contrast, a neutral antagonist like 6b-naltrexol could only

reverse the loss of LTPGABA by preventing agonist binding to the receptor. In contrast to the rapid

restoration of LTPGABA by bath application of norBNI, bath application of 6b–naltrexol did not res-

cue LTPGABA. This discrepancy cannot be explained by insufficient concentration or time of applica- tion of 6b–naltrexol, as a similar bath perfusion of 6b–naltrexol was sufficient to block the depression of EPSCs onto VTA dopamine neurons induced by bath application of the kOR agonist, U50488. The simplest explanation of our data is that acute stress induces constitutive activity of the kOR. Alternatively, norBNI may promote JNK signaling via an unknown mechanism independent of kOR receptors.

Experience-induced constitutive activity Acute stress appears to trigger a shift towards constitutively active kORs through a transient release of the endogenous kOR ligand, dynorphin (Figure 7). Our strongest evidence for this model is the

ability of the neutral antagonist 6b–naltrexol to prevent the loss of LTPGABA when administered before, but not after stress. Although 6b–naltrexol has equivalent affinity for m and k ORs,

(Wang et al., 2007) our previous work has shown that the block of LTPGABA by stress is unaffected by pre-stress administration of the mOR antagonist cyprodime (Graziane et al., 2013). Therefore,

the ability of 6b–naltrexol to prevent the stress-induced loss of LTPGABA is unlikely to involve mOR signaling and instead occurs by blocking dynorphin binding to the kOR. A single in vivo systemic

administration of the kOR agonist U50488 also blocks LTPGABA for at least five days, supporting the idea that brief agonist exposure alone is sufficient to trigger lasting constitutive kOR activity. How might activation of kORs by its endogenous ligand shift the receptor towards constitutive activity? In a heterologous cell-culture system, previous exposure to a kOR agonist alone significantly increased constitutive activity of the receptor (Wang et al., 2007). More is known regarding regulation of constitutive activity of mORs. In either cultured cells heterologously express- ing mORs (Wang et al., 1994, 2000; Liu and Prather, 2001) or in intact animals (Wang et al., 2004; Shoblock and Maidment, 2006; Meye et al., 2012), exposure to the mOR agonist trig- gers an increase in constitutive activity of mORs. Morphine-induced constitutive activity of mORs is regulated by calmodulin and protein kinases. Under basal conditions, calmodulin binding to the mORs prevents constitutive association with G-proteins. Following morphine exposure, calmodulin dissociates from the mOR, allowing constitutive activation (Wang et al., 1999, 2000). Although it is unknown whether calmodulin regulates the activity of kORs, an intricate scaffolding complex regu- lates kOR signaling (Bruchas and Chavkin, 2010), and future studies investigating the role of these

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 12 of 21 Research article Neuroscience signaling complexes in kOR activity in response to stress will be important and intriguing. Most of the work investigating constitutive activation of GPCRs has focused on enhancement of constitutive activity by administration of exogenous ligands (Meye et al., 2014), while considerably less is known about induction of constitutively active states of GPCRs by endogenous signaling. However, it was recently reported that inflammatory pain increases constitutively active mORs in the mouse spinal cord, leading to hyperalgesia and cellular dependence (Corder et al., 2013). Our data indicate that treatment with the kOR agonist U50488 alone is sufficient to produce con- stitutively active kORs on inhibitory VTA synapses. A remaining puzzle is why kOR activation by endogenous ligand can produce constitutively active receptors at some synapses but not at their neighbors, and in response to certain environmental cues (acute stress) but not to others during which dynorphin may also be released. One possibility is that receptors in different cell types may couple to different signaling cascades or scaffolding molecules. Another possibility is that coordi- nated signaling between kORs and another neurotransmitter system may be required. Our prior

work indicates that activation of glucocorticoid receptors is required for the block of LTPGABA by stress (Niehaus et al., 2010; Polter et al., 2014). Although persistent activation of these receptors is not seen after stress, it is possible that coincident activation of glucocorticoid and kappa opioid receptors leads to constitutive activation of the latter. Additionally, it has been reported that the orexin-1 receptor attenuates kOR inhibition of cAMP production, but enhances recruitment of b- arrestin and p38 MAPK activation, and both effects are prevented by the JNK inhibitor SP600125 used in our study (Robinson and McDonald, 2015). Both orexin and dynorphin are co-released from hypothalamic projections to the VTA (Chou et al., 2001; Muschamp et al., 2014; Baimel et al., 2015). This arrangement raises the possibility that release of both peptides together, or perhaps simultaneous release of dynorphin and an unknown neurotransmitter acting similarly to orexin, may initiate signaling events not triggered by dynorphin alone. The putative dual receptor signaling might be one way to induce synapse- or experience- selective constitutive kOR activity.

Regulation of LTPGABA by kORs

One unanswered question is how kORs suppress the expression of LTPGABA. Our previous studies

have shown that LTPGABA is triggered by nitric oxide-mediated activation of cGMP-protein kinase G (PKG) signaling (Nugent et al., 2007, 2009; Niehaus et al., 2010). Because an exogenous source of nitric oxide (SNAP) does not rescue potentiation following stress, the blockade is likely to occur in the presynaptic terminal between activation of guanylate cyclase and enhancement of GABAergic release. While it is possible that kOR activation generally depresses GABA release, our previous work (Graziane et al., 2013) found no change in mIPSC frequency following cold water swim stress.

These data suggest that kORs do not alter basal GABA release. Moreover, LTPGABA is also lost 24 hr after a single morphine exposure, and in this situation a cGMP analog or strong activation of sGC potentiates GABA release (Nugent et al., 2007; Niehaus et al., 2010). We therefore favor a mecha- nism by which after acute stress, constitutively-active kORs similarly act on a substrate that limits induction of plasticity without affecting basal release mechanisms, perhaps through downregulation of soluble guanylyl cyclase, or scaffolding changes that sequester PKG from its substrates.

Although it remains unknown under what conditions LTPGABA is activated in an intact animal, our

prior studies shed some light on its potential roles. LTPGABA is a heterosynaptic form of plasticity that can be triggered by a high-frequency tetanus that activates NMDAR-dependent activation of

calcium-sensitive nitric oxide synthase (Nugent et al., 2007). We therefore expect that LTPGABA would be induced when there is robust activation of excitatory inputs onto dopamine neurons.

LTPGABA may play a homeostatic role to enhance inhibition of dopamine neurons after strong

NMDAR-activating excitation. Loss of LTPGABA, therefore, would result in an imbalance between inhibitory and excitatory input onto the dopamine neuron. As GABAergic synapses on dopamine

neurons strongly control their spontaneous firing (van Zessen et al., 2012), the loss of LTPGABA is likely to prolong or enhance firing in response to salient stimuli.

Constitutive activity of kORs in the VTA and drug-seeking behavior The critical role of the VTA in reinstatement of drug seeking has been repeatedly underscored (McFarland et al., 2004; Briand et al., 2010; Graziane et al., 2013; Mantsch et al., 2016), and

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 13 of 21 Research article Neuroscience within the VTA, GABAergic synapses on VTA dopamine neurons powerfully regulate DA cell firing (Tan et al., 2012; van Zessen et al., 2012; Polter and Kauer, 2014). Our work shows that stress produces long-lasting kOR constitutive activity that is restricted to inhibitory synapses on dopamine cells, thereby affecting information stored or processed here for far longer than at excitatory synapses. We might therefore predict two phases of stress-induced kOR activation. We hypothesize that dynorphin is released during and/or immediately after stress, depressing EPSCs onto dopaminergic neurons and hyperpolarizing dopaminergic neurons, on bal- ance decreasing dopaminergic neuron excitability (Margolis et al., 2003, 2005; Ford et al., 2006). However, as dynorphin is degraded, we would expect that the strength of excitatory synapses would

return to normal levels while LTPGABA would become blocked by constitutive activity of kORs, a state lasting at least five days after swim stress. This would instead increase the firing rate of dopa- minergic neurons, particularly in response to excitatory stimuli. This increased excitability could con- tribute to the increased drive towards drug-seeking behavior upon exposure to spatial cues associated with past drug experience (i.e., return to the operant chamber), and could create a state of vulnerability to further stressors. Indeed, in rats subjected to the same cold water stress used in this study, the firing rate of dopaminergic neurons remains elevated for several days afterwards (Marinelli, 2007). Interestingly, a single dose of the kOR agonist, , has biphasic effects on reward function: immediately after administration, rats exhibit an anhedonic increase in reward thresholds to intracranial self-stimulation. However, 24 hr after salvinorin A administration, rats exhibit decreased reward thresholds, indicating an increase in reward sensitivity (Potter et al., 2011). This biphasic effect is consistent with a split between short- and long-term effects of kOR activation, perhaps due to differential mechanisms of regulation and constitutive activation of sub- sets of receptors. The circuitry of the VTA is highly complex, and dopamine neurons within the VTA exhibit physio- logical and functional heterogeneity that correlates with projection target. While disagreement remains about the most appropriate pharmacological, physiological, and anatomical markers of dif- ferent subclasses of dopamine neurons (Ford et al., 2006; Margolis et al., 2006; Lammel et al., 2008, 2011; Ungless and Grace, 2012; Baimel et al., 2017), the electrophysiological markers used here and the lateral location of our recordings within the VTA suggest to us that our population of cells may largely comprise dopamine neurons that project to the nucleus accumbens. This may be significant for drug reward, as activation of these neurons has been shown to be rewarding in mice (Lammel et al., 2012). Therefore, our study indicates that an acute stressor induces a long-lasting loss of inhibitory plasticity in circuitry that may drive rewarding behavior.

GABAergic afferents on VTA dopamine neurons can release GABA onto either GABAA or GABAB receptors. Previous studies including more recent optogenetic approaches have suggested that

GABAB receptor-targeting neurons arise from the nucleus accumbens and regulate drug-induced behaviors (Sugita et al., 1992; Cameron and Williams, 1993; McCall et al., 2017; Edwards et al.,

2017). However, our earlier work found no LTPGABA at GABAB synapses on dopamine neurons (Nugent et al., 2009), suggesting that the effects of persistently activated kORs are unlikely to involve the nucleus accumbens-VTA GABAergic afferents. Our data provide the first demonstration that constitutively active kORs in the VTA are required for stress-induced reinstatement of cocaine-seeking. Post-stress (at least 24 hr) administration of norBNI prevents reinstatement, while post-stress administration of the neutral antagonist 6b-nal- trexol does not. The ability of norBNI to modify drug-seeking behavior even when given significantly after the stressor is remarkable, and indicates the therapeutic potential of targeting kORs to reverse stress-induced neuroadaptations. The failure of 6b–naltrexol to prevent reinstatement at time points when norBNI is effective strongly suggests that the persistent increase in drug seeking induced by swim stress is mediated by constitutive activity of kORs rather than a prolonged increase in dynor- phin release. Furthermore, this result is consistent with an important role for GABAergic synapse plasticity in stress-induced drug-seeking behavior. While considerable attention has been given to the role of LTP at excitatory synapses in the VTA, the kOR block by norBNI does not prevent stress from potentiating excitatory synapses on dopamine neurons (Graziane et al., 2013). Our current work confirms that the loss of LTP at GABAergic synapses in the VTA is highly correlated with stress- induced drug-seeking.

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 14 of 21 Research article Neuroscience kORs as targets for substance abuse kORs have shown promise as a potential drug target for substance use and mood disorders (Bruchas et al., 2010; Van’t Veer and Carlezon, 2013; Crowley and Kash, 2015) and our work suggests a novel way in which kOR signaling may go awry. In preclinical models, kOR antagonists have shown potential efficacy for depression and for compulsive and stress-induced drug use (Bruchas et al., 2010; Wee and Koob, 2010). Numerous clinical trials are in progress using kOR ligands to target substance use disorders and depression (Ehrich et al., 2015; Karp et al., 2014; Chavkin and Koob, 2016; Ling et al., 2016; Nasser et al., 2016). However, many of these trials use , a partial agonist at kORs, or novel compounds which may lack inverse agonist activ- ity, neither of which would reduce activity of constitutive kORs (Karp et al., 2014; Rorick- Kehn et al., 2014). Our study suggests that future drug development should consider excess kOR activity through receptor signaling as well as at the level of ligand binding. Similarly, disappointing results or minimal effects in clinical trials may not represent failure of kORs as a pharmaceutical target, but a need to consider drugs that target specific conformations of the kOR that promote constitutive signaling. An alternative strategy would be to target JNK signaling in the VTA, as norBNI appears to rescue kOR function by activating JNK. An intriguing implication of our studies comes from our data that consti- tutive activity of kORs at inhibitory synapses in the VTA lasts only five to ten days following acute stress (Polter et al., 2014), a considerably shorter time period than the 14–21 days typical for turn- over of kORs (McLaughlin et al., 2004). This suggests that constitutive activity of the kORs may be terminated by an unidentified active mechanism. Future studies investigating such a mechanism could identify targets that could be recruited to promote resilience to stress. Our work demonstrates a novel mechanism of experience-dependent regulation of kORs, and highlights the ability of modu- lation of kORs to reverse stress-induced neuroadaptations and behavioral deficiencies well after the stressor has occurred. Further study of the mechanisms of constitutive activation of kORs may yield numerous potential targets for the treatment of substance use disorders and other stress-linked illnesses.

Materials and methods Animals All procedures were carried out in accordance with the guidelines of the National Institutes of Health for animal care and use, and were approved by the Brown University Institutional Animal Care and Use Committee or by the University of Wyoming Institutional Animal Care and Use Committee. For slice electrophysiology studies, male and female Sprague-Dawley rats (P16-25) were maintained on a 12 hr light / dark cycle and provided food and water ad libitum. For self-administration studies, male Sprague-Dawley rats (350–450g) were bred in-house and individually housed in a temperature-con- trolled room with a 12 hr reverse light/dark cycle. All animals were given ad libitum access to water throughout experimentation, except during times in which they were in the operant chambers (described below). Rats were 60–70 days old at the start of behavioral experiments.

Acute forced swim stress Stress was administered by a modified Porsolt forced swim task (Niehaus et al., 2010). Rats were placed for 5 min in cold water (4–6˚C), then dried and allowed to recover in a warmed cage for two hours before returning to the home cage. U50488 (5 mg/kg) and 6b-naltrexol (10 mg/kg) were dis- solved in saline or 10% DMSO in saline, respectively. Vehicle-injected animals were given an injection of the equivalent volume. For some experiments, animals given vehicle injections at varying time points were collapsed into a single group. Brain slices were prepared at several time points after stress exposure, as described below.

Preparation of brain slices Horizontal midbrain slices (250 mm) were prepared as previously described from deeply anesthetized Sprague-Dawley rats (Nugent et al., 2007; Niehaus et al., 2010; Polter et al., 2014). Slices were stored for at least 1 hr at 34˚C in oxygenated HEPES holding solution (in mM): 86 NaCl, 2.5 KCl, 1.2 NaH2PO4, 35 NaHCO3, 20 HEPES, 25 glucose, 5 sodium ascorbate, 2 thiourea, 3 sodium pyruvate, 1

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 15 of 21 Research article Neuroscience

MgSO4.7H2O, 2 CaCl2.2H2O(Ting et al., 2014). Slices were then transferred to a recording chamber

where they were submerged in ACSF containing (in mM): 126 NaCl, 21.4 NaHCO3, 2.5 KCl, 1.2

NaH2PO4, 2.4 CaCl2, 1.0 MgSO4, 11.1 glucose. Electrophysiology General methods were as previously reported (Niehaus et al., 2010; Polter et al., 2014). Midbrain slices were continuously perfused at 1.5–2 mL / min. Patch pipettes were filled with (in mM): 125

KCl, 2.8 NaCl, 2 MgCl2, 2 ATP-Na+, 0.3 GTP-Na+, 0.6 EGTA, and 10 HEPES. To record IPSCs, the extracellular solution was ACSF (28–32˚C) containing: 6,7-dinitroquinoxaline- 2,3-dione (DNQX; 10 mM) and strychnine (1 mM), to block AMPA and glycine receptors respectively. To record EPSCs, 100 mM picrotoxin was added to the ACSF. Dopaminergic neurons, which comprise about 70% of all

VTA neurons, were identified by the presence of a large Ih (>50 pA) during a voltage step from À50

mV to À100 mV. GABAA receptor-mediated IPSCs were stimulated using a bipolar stainless steel stimulating electrode placed 100–300 mm rostral to the recording site in the VTA. Cells were volt- age-clamped at À70 mV and input resistance and series resistance were monitored throughout the experiment; cells were discarded if these values changed by more than 15% during the experiment.

NO-triggered LTP 3-isobutyl-1-methylxanthine (IBMX; 100 mM) was used to inhibit phosphodiesterase-mediated degra-

dation of cGMP and applied via perfused ACSF for at least 10 min prior to induction of LTPGABA by application of the NO donor, SNAP (S-nitroso-N-acetylpenicillamine, 400 mM). Control animals (vehi- cle-injected stressed or unstressed animals) were interleaved with experimental animals (drug- injected stressed animals). Where indicated, NorBNI (100 nM), 6b-naltrexol (10 mM), and SP600135

(20 mM) were bath applied to slices at least 10 min prior to induction of LTPGABA. Self-administration Rats were anesthetized with HCl (87 mg/kg, i.m.) and xylazine (13 mg/kg, i.m.) and implanted with intravenous jugular catheters. In order to protect against infection and maintain cath- eter patency, catheters were flushed daily with 0.2 mL of a mixed cefazolin (0.1 gm/ml) and heparin (100 IU) solution. Rats were allowed to recover for one week before behavioral testing. All self- administration procedures were conducted in standard operant chambers (Med Associates, St. Albans, VT; 30.5 cm x 24.1 cm x 21.0 cm). Each box contained a house light (illuminated throughout behavioral testing) two retractable levers, a cue light, and tone generator. Prior to beginning cocaine self-administration training, animals were food deprived for 24 hr and subsequently placed into the operant chambers overnight for 14 hr. During this session, a response to the active lever (the left lever) resulted in the delivery of a single 45 mg food pellet (#F0165, Bio-Serv, Flemington, NJ) and the presentation of a compound cue (illumination of light above the active lever +5 s tone, 2900 Hz), followed by a 25 s timeout period. Responses to the inactive lever (the right lever) had no pro- grammed consequences but were recorded. Total rewards received were also recorded. On the next day, cocaine self-administration training began. During this time, a response to the active lever yielded a 0.05 ml infusion of 0.20 mg of cocaine (dissolved in 0.9% saline) as well as the presentation of the compound cue (light + tone). Self-administration continued (2 hr/daily) until animals reliably pressed the active lever (3d with minimum of 10 cocaine infusions received). Following the acquisi- tion of cocaine self-administration, all animals underwent extinction training, during which responses to the previously active lever yielded the compound cue but no longer produced drug infusion. Ani- mals were food-restricted to 80% of their body weight during self-administration training. During extinction, animals were given ad libitum access to food in the home cage. Extinction procedures continued until animals reached extinction criteria (3d with less than 10 active lever presses).

Forced swim stress and reinstatement The day following the last extinction session, rats were subjected to a 5 min forced swim stress in cold water (4-6˚C, Saal et al., 2003; Niehaus et al., 2010). Rats were then split into three groups, receiving (i.p.) injections of either saline (1 ml/kg), norBNI (10 mg/kg), or 6b-Naltrexol (10 mg/kg). 24 hr after swim stress, rats in the saline and norBNI groups were injected and left undisturbed in their home cage for one day. Rats in the 6b-naltrexol group were injected one hour prior to reinstatement

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 16 of 21 Research article Neuroscience testing. At 48 hr after swim stress, all animals were subjected to reinstatement testing, which simi- larly to extinction yielded the compound cue but no drug infusion.

Analysis Magnitude of LTP was determined as mean IPSC amplitude for 5 min just before application of SNAP compared with mean IPSC amplitude from 10–15 min after SNAP application, unless other- wise noted. Data are presented as means ± SEM of the percent IPSC amplitude normalized to IPSCs in the 10 min prior to SNAP application. Statistical methods were not used to determine sample size. Sample size was based on our prior experience and previously published studies (Graziane et al., 2013; Polter et al., 2014). All reported n’s are the number of animals (biological replicates), unless otherwise noted. Significance was determined using a two-tailed Student’s t-test or a one-way ANOVA with a significance level of p<0.05. All post-hoc comparisons were done using Dunnett’s test unless otherwise noted. Self-administration data were analyzed using paired t-tests.

Materials IBMX was obtained from Enzo Life Sciences. NorBNI, U50488, and SNAP were obtained from Tocris Biosciences. DNQX, picrotoxin, strychnine, and 6b-naltrexol were obtained from Sigma-Aldrich. SP600125 was obtained from Calbiochem.

Acknowledgements The authors would like to thank Dr. Jennifer Whistler for helpful discussion of kOR antagonists, Dr. Michael Bruchas for the suggestion that we test the role of JNK, and Ayumi Tsuda and Elodi Healy for technical assistance.

Additional information

Funding Funder Grant reference number Author National Institute on Drug R01DA011289 Julie A Kauer Abuse Brain and Behavior Research Young Investigator Award Abigail M Polter Foundation National Institute of Mental K99MH106757 Abigail M Polter Health National Institute on Drug R01DA040965 Travis E Brown Abuse National Institute of General P30 GM 103398-32128 Travis E Brown Medical Sciences

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Author contributions AMP, Conceptualization, Formal analysis, Supervision, Funding acquisition, Investigation, Methodol- ogy, Writing—original draft, Project administration, Writing—review and editing; KB, RWC, PMD, Formal analysis, Investigation, Writing—review and editing; NMG, Investigation, Writing—review and editing; TEB, Supervision, Funding acquisition, Investigation, Writing—review and editing; JAK, Conceptualization, Formal analysis, Supervision, Funding acquisition, Writing—original draft, Project administration, Writing—review and editing

Author ORCIDs Abigail M Polter, http://orcid.org/0000-0003-0151-0996 Julie A Kauer, http://orcid.org/0000-0002-3362-1642

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 17 of 21 Research article Neuroscience Ethics Animal experimentation: All procedures were carried out in accordance with the guidelines of the National Institutes of Health for animal care and use, and were approved by the Brown University Institutional Animal Care and Use Committee ( protocol #1411000106) or by the University of Wyom- ing Institutional Animal Care and Use Committee (protocol # #20150909TB00195-92).

References Baimel C, Bartlett SE, Chiou LC, Lawrence AJ, Muschamp JW, Patkar O, Tung LW, Borgland SL. 2015. Orexin/ hypocretin role in reward: implications for opioid and other addictions. British Journal of Pharmacology 172: 334–348. doi: 10.1111/bph.12639, PMID: 24641197 Baimel C, Lau BK, Qiao M, Borgland SL. 2017. Projection-Target-Defined effects of orexin and dynorphin on VTA dopamine neurons. Cell Reports 18:1346–1355. doi: 10.1016/j.celrep.2017.01.030, PMID: 28178514 Becker JA, Wallace A, Garzon A, Ingallinella P, Bianchi E, Cortese R, Simonin F, Kieffer BL, Pessi A. 1999. Ligands for kappa-opioid and ORL1 receptors identified from a conformationally constrained peptide combinatorial library. Journal of Biological Chemistry 274:27513–27522. doi: 10.1074/jbc.274.39.27513, PMID: 10488086 Briand LA, Vassoler FM, Pierce RC, Valentino RJ, Blendy JA. 2010Ventral tegmental afferents in stress-induced reinstatement: the role of cAMP response element-binding protein. Journal of Neuroscience 30:16149–16159. doi: 10.1523/JNEUROSCI.2827-10.2010, PMID: 21123561 Bruchas MR, Chavkin C. 2010. Kinase cascades and ligand-directed signaling at the kappa opioid receptor. Psychopharmacology 210:137–147. doi: 10.1007/s00213-010-1806-y, PMID: 20401607 Bruchas MR, Land BB, Chavkin C. 2010. The dynorphin/kappa opioid system as a modulator of stress-induced and pro-addictive behaviors. Brain Research 1314:44–55. doi: 10.1016/j.brainres.2009.08.062, PMID: 19716811 Bruchas MR, Yang T, Schreiber S, Defino M, Kwan SC, Li S, Chavkin C. 2007. Long-acting kappa opioid antagonists disrupt receptor signaling and produce noncompetitive effects by activating c-Jun N-terminal kinase. Journal of Biological Chemistry 282:29803–29811. doi: 10.1074/jbc.M705540200, PMID: 17702750 Cameron DL, Williams JT. 1993. Dopamine D1 receptors facilitate transmitter release. Nature 366:344–347. doi: 10.1038/366344a0, PMID: 8247128 Chavkin C, Koob GF. 2016. Dynorphin, dysphoria, and dependence: the stress of addiction. Neuropsychopharmacology 41:373–374. doi: 10.1038/npp.2015.258, PMID: 26657953 Chen BT, Bowers MS, Martin M, Hopf FW, Guillory AM, Carelli RM, Chou JK, Bonci A. 2008. Cocaine but not natural reward self-administration nor passive cocaine infusion produces persistent LTP in the VTA. Neuron 59: 288–297. doi: 10.1016/j.neuron.2008.05.024, PMID: 18667156 Chiu TT, Yung LY, Wong YH. 1996. Inverse agonistic effect of ICI-174,864 on the cloned delta-opioid receptor: role of G protein and adenylyl cyclase activation. Molecular Pharmacology 50:1651–1657. PMID: 8967989 Chou TC, Lee CE, Lu J, Elmquist JK, Hara J, Willie JT, Beuckmann CT, Chemelli RM, Sakurai T, Yanagisawa M, Saper CB, Scammell TE. 2001. Orexin (hypocretin) neurons contain dynorphin. Journal of Neuroscience 21: RC168. PMID: 11567079 Conrad KL, McCutcheon JE, Cotterly LM, Ford KA, Beales M, Marinelli M. 2010. Persistent increases in cocaine- seeking behavior after acute exposure to cold swim stress. Biological Psychiatry 68:303–305. doi: 10.1016/j. biopsych.2010.03.030, PMID: 20494337 Corder G, Doolen S, Donahue RR, Winter MK, Jutras BL, He Y, Hu X, Wieskopf JS, Mogil JS, Storm DR, Wang ZJ, McCarson KE, Taylor BK. 2013. Constitutive m-opioid receptor activity leads to long-term endogenous analgesia and dependence. Science 341:1394–1399. doi: 10.1126/science.1239403, PMID: 24052307 Crowley NA, Kash TL. 2015. Kappa opioid receptor signaling in the brain: circuitry and implications for treatment. Progress in Neuro-Psychopharmacology and Biological Psychiatry 62:51–60. doi: 10.1016/j.pnpbp. 2015.01.001, PMID: 25592680 Edwards NJ, Tejeda HA, Pignatelli M, Zhang S, McDevitt RA, Wu J, Bass CE, Bettler B, Morales M, Bonci A. 2017. Circuit specificity in the inhibitory architecture of the VTA regulates cocaine-induced behavior. Nature Neuroscience 20:438–448. doi: 10.1038/nn.4482, PMID: 28114294 Ehrich E, Turncliff R, Du Y, Leigh-Pemberton R, Fernandez E, Jones R, Fava M. 2015. Evaluation of opioid modulation in major depressive disorder. Neuropsychopharmacology 40:1448–1455. doi: 10.1038/npp.2014. 330, PMID: 25518754 Ehrich JM, Messinger DI, Knakal CR, Kuhar JR, Schattauer SS, Bruchas MR, Zweifel LS, Kieffer BL, Phillips PE, Chavkin C. 2015. Kappa opioid Receptor-Induced aversion requires p38 MAPK activation in VTA dopamine neurons. Journal of Neuroscience 35:12917–12931. doi: 10.1523/JNEUROSCI.2444-15.2015, PMID: 26377476 Endoh T, Matsuura H, Tanaka C, Nagase H. 1992. Nor-binaltorphimine: a potent and selective kappa-opioid with long-lasting activity in vivo. Archives Internationales De Pharmacodynamie Et De Therapie 316:30–42. PMID: 1326932 Ford CP, Beckstead MJ, Williams JT. 2007. Kappa opioid inhibition of somatodendritic dopamine inhibitory postsynaptic currents. Journal of Neurophysiology 97:883–891. doi: 10.1152/jn.00963.2006, PMID: 17122312 Ford CP, Mark GP, Williams JT. 2006. Properties and opioid inhibition of mesolimbic dopamine neurons vary according to target location. Journal of Neuroscience 26:2788–2797. doi: 10.1523/JNEUROSCI.4331-05.2006, PMID: 16525058

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 18 of 21 Research article Neuroscience

Goeders NE, Guerin GF. 1994. Non-contingent electric footshock facilitates the acquisition of intravenous cocaine self-administration in rats. Psychopharmacology 114:63–70. doi: 10.1007/BF02245445, PMID: 7846208 Graziane NM, Polter AM, Briand LA, Pierce RC, Kauer JA. 2013. Kappa opioid receptors regulate stress-induced cocaine seeking and synaptic plasticity. Neuron 77:942–954. doi: 10.1016/j.neuron.2012.12.034, PMID: 23473323 Haney M, Maccari S, Le Moal M, Simon H, Piazza PV. 1995. Social stress increases the acquisition of cocaine self- administration in male and female rats. Brain Research 698:46–52. doi: 10.1016/0006-8993(95)00788-R, PMID: 8581502 In˜ iguez SD, Vialou V, Warren BL, Cao JL, Alcantara LF, Davis LC, Manojlovic Z, Neve RL, Russo SJ, Han MH, Nestler EJ, Bolan˜ os-Guzma´n CA. 2010. Extracellular signal-regulated kinase-2 within the ventral tegmental area regulates responses to stress. Journal of Neuroscience 30:7652–7663. doi: 10.1523/JNEUROSCI.0951-10.2010, PMID: 20519540 Karp JF, Butters MA, Begley AE, Miller MD, Lenze EJ, Blumberger DM, Mulsant BH, Reynolds CF. 2014. Safety, tolerability, and clinical effect of low-dose buprenorphine for treatment-resistant depression in midlife and older adults. The Journal of Clinical Psychiatry 75:e785–e793. doi: 10.4088/JCP.13m08725, PMID: 25191915 Kauer JA, Malenka RC. 2007. Synaptic plasticity and addiction. Nature Reviews. Neuroscience 8:844–858. doi: 10.1038/nrn2234, PMID: 17948030 Lammel S, Hetzel A, Ha¨ ckel O, Jones I, Liss B, Roeper J. 2008. Unique properties of mesoprefrontal neurons within a dual mesocorticolimbic dopamine system. Neuron 57:760–773. doi: 10.1016/j.neuron.2008.01.022, PMID: 18341995 Lammel S, Ion DI, Roeper J, Malenka RC. 2011. Projection-specific modulation of dopamine neuron synapses by aversive and rewarding stimuli. Neuron 70:855–862. doi: 10.1016/j.neuron.2011.03.025, PMID: 21658580 Lammel S, Lim BK, Ran C, Huang KW, Betley MJ, Tye KM, Deisseroth K, Malenka RC. 2012. Input-specific control of reward and aversion in the ventral tegmental area. Nature 491:212–217. doi: 10.1038/nature11527, PMID: 23064228 Land BB, Bruchas MR, Schattauer S, Giardino WJ, Aita M, Messinger D, Hnasko TS, Palmiter RD, Chavkin C. 2009. Activation of the kappa opioid receptor in the dorsal raphe nucleus mediates the aversive effects of stress and reinstates drug seeking. PNAS 106:19168–19173. doi: 10.1073/pnas.0910705106, PMID: 19864633 Ling W, Hillhouse MP, Saxon AJ, Mooney LJ, Thomas CM, Ang A, Matthews AG, Hasson A, Annon J, Sparenborg S, Liu DS, McCormack J, Church S, Swafford W, Drexler K, Schuman C, Ross S, Wiest K, Korthuis PT, Lawson W, et al. 2016. buprenorphine + plus for the treatment of cocaine dependence: the Cocaine Use Reduction with Buprenorphine (CURB) study. Addiction 111:1416–1427. doi: 10. 1111/add.13375, PMID: 26948856 Liu JG, Prather PL. 2001. Chronic exposure to mu-opioid agonists produces constitutive activation of mu-opioid receptors in direct proportion to the efficacy of the agonist used for pretreatment. Molecular Pharmacology 60:53–62. PMID: 11408600 Mantsch JR, Baker DA, Funk D, LeˆAD, Shaham Y. 2016. Stress-Induced reinstatement of drug seeking: 20 years of progress. Neuropsychopharmacology : Official Publication of the American College of Neuropsychopharmacology 41:1–22. doi: 10.1038/npp.2015.142, PMID: 25976297 Margolis EB, Hjelmstad GO, Bonci A, Fields HL. 2003. Kappa-opioid agonists directly inhibit midbrain dopaminergic neurons. Journal of Neuroscience 23:9981–9986. PMID: 14602811 Margolis EB, Hjelmstad GO, Bonci A, Fields HL. 2005. Both kappa and mu opioid agonists inhibit glutamatergic input to ventral tegmental area neurons. Journal of Neurophysiology 93:3086–3093. doi: 10.1152/jn.00855. 2004, PMID: 15615834 Margolis EB, Lock H, Chefer VI, Shippenberg TS, Hjelmstad GO, Fields HL. 2006. Kappa selectively control dopaminergic neurons projecting to the prefrontal cortex. PNAS 103:2938–2942. doi: 10.1073/pnas. 0511159103, PMID: 16477003 Marinelli M. 2007. Dopaminergic reward pathways and effects of stress. In: Stress and Addiction. p. 41–83. doi: 10.1016/B978-012370632-4/50006-1 McCall NM, Kotecki L, Dominguez-Lopez S, Marron Fernandez de Velasco E, Carlblom N, Sharpe AL, Beckstead MJ, Wickman K. 2017. Selective ablation of GIRK channels in dopamine neurons alters behavioral effects of cocaine in mice. Neuropsychopharmacology 42:707–715. doi: 10.1038/npp.2016.138, PMID: 27468917 McFarland K, Davidge SB, Lapish CC, Kalivas PW. 2004. Limbic and motor circuitry underlying footshock- induced reinstatement of cocaine-seeking behavior. Journal of Neuroscience 24:1551–1560. doi: 10.1523/ JNEUROSCI.4177-03.2004, PMID: 14973230 McLaughlin JP, Marton-Popovici M, Chavkin C. 2003. Kappa opioid receptor antagonism and gene disruption block stress-induced behavioral responses. Journal of Neuroscience 23. PMID: 12843270 McLaughlin JP, Myers LC, Zarek PE, Caron MG, Lefkowitz RJ, Czyzyk TA, Pintar JE, Chavkin C. 2004Prolonged kappa opioid receptor phosphorylation mediated by G-protein receptor kinase underlies sustained tolerance. Journal of Biological Chemistry 279:1810–1818. doi: 10.1074/jbc.M305796200, PMID: 14597630 Melief EJ, Miyatake M, Bruchas MR, Chavkin C. 2010. Ligand-directed c-Jun N-terminal kinase activation disrupts opioid receptor signaling. PNAS 107:11608–11613. doi: 10.1073/pnas.1000751107, PMID: 20534436 Melief EJ, Miyatake M, Carroll FI, Be´guin C, Carlezon WA, Cohen BM, Grimwood S, Mitch CH, Rorick-Kehn L, Chavkin C. 2011. Duration of action of a broad range of selective k-opioid receptor antagonists is positively correlated with c-Jun N-terminal kinase-1 activation. Molecular Pharmacology 80:920–929. doi: 10.1124/mol. 111.074195, PMID: 21832171

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 19 of 21 Research article Neuroscience

Meye FJ, Ramakers GM, Adan RA. 2014.) ‘ The vital role of constitutive GPCR activity in the mesolimbic dopamine system. Translational Psychiatry 4:e361. doi: 10.1038/tp.2013.130, PMID: 24518399 Meye FJ, van Zessen R, Smidt MP, Adan RA, Ramakers GM. 2012. Morphine withdrawal enhances constitutive m- opioid receptor activity in the ventral tegmental area. Journal of Neuroscience 32:16120–16128. doi: 10.1523/ JNEUROSCI.1572-12.2012, PMID: 23152596 Muschamp JW, Hollander JA, Thompson JL, Voren G, Hassinger LC, Onvani S, Kamenecka TM, Borgland SL, Kenny PJ, Carlezon WA. 2014. Hypocretin (orexin) facilitates reward by attenuating the antireward effects of its cotransmitter dynorphin in ventral tegmental area. PNAS 111:E1648–E1655. doi: 10.1073/pnas.1315542111, PMID: 24706819 Nasser AF, Greenwald MK, Vince B, Fudala PJ, Twumasi-Ankrah P, Liu Y, Jones JP, Heidbreder C. 2016. Sustained-Release buprenorphine (RBP-6000) Blocks the effects of opioid challenge with in subjects with . Journal of Clinical Psychopharmacology 36:18–26. doi: 10.1097/JCP. 0000000000000434, PMID: 26650971 Niehaus JL, Murali M, Kauer JA. 2010Drugs of abuse and stress impair LTP at inhibitory synapses in the ventral tegmental area. European Journal of Neuroscience 32:108–117. doi: 10.1111/j.1460-9568.2010.07256.x, PMID: 20608969 Nugent FS, Niehaus JL, Kauer JA. 2009. PKG and PKA signaling in LTP at GABAergic synapses. Neuropsychopharmacology 34:1829–1842. doi: 10.1038/npp.2009.5, PMID: 19194373 Nugent FS, Penick EC, Kauer JA. 2007. Opioids block long-term potentiation of inhibitory synapses. Nature 446: 1086–1090. doi: 10.1038/nature05726, PMID: 17460674 Piazza PV, Deminiere JM, le Moal M, Simon H. 1990. Stress- and pharmacologically-induced behavioral sensitization increases vulnerability to acquisition of amphetamine self-administration. Brain Research 514:22– 26. doi: 10.1016/0006-8993(90)90431-A, PMID: 2357527 Polter AM, Bishop RA, Briand LA, Graziane NM, Pierce RC, Kauer JA. 2014. Poststress block of kappa opioid receptors rescues long-term potentiation of inhibitory synapses and prevents reinstatement of cocaine seeking. Biological Psychiatry 76:785–793. doi: 10.1016/j.biopsych.2014.04.019, PMID: 24957331 Polter AM, Kauer JA. 2014. Stress and VTA synapses: implications for addiction and depression. European Journal of Neuroscience 39:1179–1188. doi: 10.1111/ejn.12490, PMID: 24712997 Potter DN, Damez-Werno D, Carlezon WA, Cohen BM, Chartoff EH. 2011. Repeated exposure to the k-opioid receptor agonist salvinorin A modulates extracellular signal-regulated kinase and reward sensitivity. Biological Psychiatry 70:744–753. doi: 10.1016/j.biopsych.2011.05.021, PMID: 21757186 Raehal KM, Lowery JJ, Bhamidipati CM, Paolino RM, Blair JR, Wang D, Sade´e W, Bilsky EJ. 2005. In vivo characterization of 6beta-naltrexol, an opioid ligand with less inverse agonist activity compared with naltrexone and naloxone in opioid-dependent mice. Journal of Pharmacology and Experimental Therapeutics 313:1150– 1162. doi: 10.1124/jpet.104.082966, PMID: 15716384 Ramsey NF, Van Ree JM. 1993. Emotional but not physical stress enhances intravenous cocaine self- administration in drug-naive rats. Brain Research 608:216–222. doi: 10.1016/0006-8993(93)91461-Z, PMID: 84 95356 Redila VA, Chavkin C. 2008. Stress-induced reinstatement of cocaine seeking is mediated by the kappa opioid system. Psychopharmacology 200:59–70. doi: 10.1007/s00213-008-1122-y, PMID: 18575850 Robinson JD, McDonald PH. 2015. The orexin 1 receptor modulates kappa opioid receptor function via a JNK- dependent mechanism. Cellular Signalling 27:1449–1456. doi: 10.1016/j.cellsig.2015.03.026, PMID: 25857454 Rogala B, Li Y, Li S, Chen X, Kirouac GJ. 2012. Effects of a post-shock injection of the kappa opioid receptor antagonist (norBNI) on fear and anxiety in rats. PLoS One 7:e49669. doi: 10.1371/journal. pone.0049669, PMID: 23166745 Rorick-Kehn LM, Witkin JM, Statnick MA, Eberle EL, McKinzie JH, Kahl SD, Forster BM, Wong CJ, Li X, Crile RS, Shaw DB, Sahr AE, Adams BL, Quimby SJ, Diaz N, Jimenez A, Pedregal C, Mitch CH, Knopp KL, Anderson WH, et al. 2014. LY2456302 is a novel, potent, orally-bioavailable small molecule kappa-selective antagonist with activity in animal models predictive of efficacy in mood and addictive disorders. Neuropharmacology 77:131– 144. doi: 10.1016/j.neuropharm.2013.09.021, PMID: 24071566 Russell SE, Rachlin AB, Smith KL, Muschamp J, Berry L, Zhao Z, Chartoff EH. 2014. Sex differences in sensitivity to the depressive-like effects of the kappa opioid receptor agonist U-50488 in rats. Biological Psychiatry 76: 213–222. doi: 10.1016/j.biopsych.2013.07.042, PMID: 24090794 Saal D, Dong Y, Bonci A, Malenka RC. 2003. Drugs of abuse and stress trigger a common synaptic adaptation in dopamine neurons. Neuron 37:577–582. doi: 10.1016/S0896-6273(03)00021-7, PMID: 12597856 Sade´ e W, Wang D, Bilsky EJ. 2005. Basal opioid receptor activity, neutral antagonists, and therapeutic opportunities. Life Sciences 76:1427–1437. doi: 10.1016/j.lfs.2004.10.024, PMID: 15680308 Seifert R, Wenzel-Seifert K. 2002. Constitutive activity of G-protein-coupled receptors: cause of disease and common property of wild-type receptors. Naunyn-Schmiedeberg’s Archives of Pharmacology 366:381–416. doi: 10.1007/s00210-002-0588-0, PMID: 12382069 Shaham Y, Rodaros D, Stewart J. 1994. Reinstatement of -reinforced behavior following long-term extinction: implications for the treatment of relapse to drug taking. Behavioural Pharmacology 5:360–364. doi: 10.1097/00008877-199406000-00015, PMID: 11224287 Shaham Y, Stewart J. 1994. Exposure to mild stress enhances the reinforcing efficacy of intravenous heroin self- administration in rats. Psychopharmacology 114:523–527. doi: 10.1007/BF02249346, PMID: 7855213 Shaham Y, Stewart J. 1995. Stress reinstates heroin-seeking in drug-free animals: an effect mimicking heroin, not withdrawal. Psychopharmacology 119:334–341. doi: 10.1007/BF02246300, PMID: 7675970

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 20 of 21 Research article Neuroscience

Shoblock JR, Maidment NT. 2006. Constitutively active micro opioid receptors mediate the enhanced conditioned aversive effect of naloxone in morphine-dependent mice. Neuropsychopharmacology : Official Publication of the American College of Neuropsychopharmacology 31:171–177. doi: 10.1038/sj.npp.1300782, PMID: 15956992 Sirohi S, Walker BM. 2015. Maturational alterations in constitutive activity of medial prefrontal cortex kappa- opioid receptors in Wistar rats. Journal of Neurochemistry 135:659–665. doi: 10.1111/jnc.13279, PMID: 26257334 Sugita S, Johnson SW, North RA. 1992. Synaptic inputs to GABAA and GABAB receptors originate from discrete afferent neurons. Neuroscience Letters 134:207–211. doi: 10.1016/0304-3940(92)90518-C, PMID: 1350333 Tan KR, Yvon C, Turiault M, Mirzabekov JJ, Doehner J, Laboue`be G, Deisseroth K, Tye KM, Lu¨ scher C. 2012. GABA neurons of the VTA drive conditioned place aversion. Neuron 73:1173–1183. doi: 10.1016/j.neuron. 2012.02.015, PMID: 22445344 Ting JT, Daigle TL, Chen Q, Feng G. 2014. Acute brain slice methods for adult and aging animals: application of targeted patch clamp analysis and optogenetics. Methods in Molecular Biology 1183:1–21. doi: 10.1007/978-1- 4939-1096-0_14, PMID: 25023312 Ungless MA, Grace AA. 2012. Are you or aren’t you? challenges associated with physiologically identifying dopamine neurons. Trends in Neurosciences 35:422–430. doi: 10.1016/j.tins.2012.02.003, PMID: 22459161 Ungless MA, Whistler JL, Malenka RC, Bonci A. 2001. Single cocaine exposure in vivo induces long-term potentiation in dopamine neurons. Nature 411:583–587. doi: 10.1038/35079077, PMID: 11385572 van Zessen R, Phillips JL, Budygin EA, Stuber GD. 2012. Activation of VTA GABA neurons disrupts reward consumption. Neuron 73:1184–1194. doi: 10.1016/j.neuron.2012.02.016, PMID: 22445345 Van’t Veer A, Carlezon WA. 2013. Role of kappa-opioid receptors in stress and anxiety-related behavior. Psychopharmacology 229:435–452. doi: 10.1007/s00213-013-3195-5, PMID: 23836029 Wang D, Raehal KM, Lin ET, Lowery JJ, Kieffer BL, Bilsky EJ, Sade´e W. 2004. Basal signaling activity of mu opioid receptor in mouse brain: role in dependence. Journal of Pharmacology and Experimental Therapeutics 308:512–520. doi: 10.1124/jpet.103.054049, PMID: 14600246 Wang D, Sade´e W, Quillan JM. 1999Calmodulin binding to G protein-coupling domain of opioid receptors. Journal of Biological Chemistry 274:22081–22088. doi: 10.1074/jbc.274.31.22081, PMID: 10419536 Wang D, Sun X, Sadee W. 2007. Different effects of opioid antagonists on mu-, delta-, and kappa-opioid receptors with and without agonist pretreatment. Journal of Pharmacology and Experimental Therapeutics 321:544–552. doi: 10.1124/jpet.106.118810, PMID: 17267582 Wang D, Surratt CK, Sade´e W. 2000. Calmodulin regulation of basal and agonist-stimulated G protein coupling by the mu-opioid receptor (OP(3)) in morphine-pretreated cell. Journal of Neurochemistry 75:763–771. doi: 10. 1046/j.1471-4159.2000.0750763.x, PMID: 10899953 Wang Z, Bilsky EJ, Porreca F, Sade´e W. 1994. Constitutive mu opioid receptor activation as a regulatory mechanism underlying narcotic tolerance and dependence. Life Sciences 54:PL339–PL350. doi: 10.1016/0024- 3205(94)90022-1, PMID: 7514710 Wee S, Koob GF. 2010. The role of the dynorphin-kappa opioid system in the reinforcing effects of drugs of abuse. Psychopharmacology 210:121–135. doi: 10.1007/s00213-010-1825-8, PMID: 20352414 Young SR, Chuang SC, Zhao W, Wong RK, Bianchi R. 2013. Persistent receptor activity underlies group I mGluR- mediated cellular plasticity in CA3 neuron. Journal of Neuroscience 33:2526–2540. doi: 10.1523/JNEUROSCI. 3338-12.2013, PMID: 23392681 Zhou Y, Leri F, Grella SL, Aldrich JV, Kreek MJ. 2013. Involvement of dynorphin and kappa opioid receptor in yohimbine-induced reinstatement of heroin seeking in rats. Synapse 67:358–361. doi: 10.1002/syn.21638, PMID: 23345075

Polter et al. eLife 2017;6:e23785. DOI: 10.7554/eLife.23785 21 of 21